How do we make decisions to maximize positive outcomes? Elucidating the neural mechanisms that support this ability may lead to fundamental insights into human behavior and is crucial for the treatment of psychiatric disorders that feature maladaptive decisions (Maia & Frank, 2011). Animal studies suggest that dopaminergic neurons in the substantia nigra guide reinforcement learning, however, because animal studies typically examine behavioral adaptation following primary rewards, it is unclear how these studies generalize to human behavior, which is often motivated by higher-order rational and social rewards. We propose to study neural activity in the BG of patients undergoing surgery for the implantation of a deep brain stimulator (DBS) device for the treatment of Parkinson's Disease (PD). We will record and enhance putative dopaminergic activity in the human substantia nigra (via electrical stimulation) as participants engage in a probability learning task with abstract, audio-visual feedback. Through the combined use of microelectrode recordings, microstimulation and computational modeling, we seek to relate the phasic activity of dopaminergic activity in the human SN with reinforcement learning.